curl/lib/conncache.h

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

202 lines
7.3 KiB
C
Raw Normal View History

#ifndef HEADER_CURL_CONNCACHE_H
#define HEADER_CURL_CONNCACHE_H
/***************************************************************************
* _ _ ____ _
* Project ___| | | | _ \| |
* / __| | | | |_) | |
* | (__| |_| | _ <| |___
* \___|\___/|_| \_\_____|
*
* Copyright (C) Daniel Stenberg, <daniel@haxx.se>, et al.
* Copyright (C) Linus Nielsen Feltzing, <linus@haxx.se>
*
* This software is licensed as described in the file COPYING, which
* you should have received as part of this distribution. The terms
2020-11-04 21:02:01 +08:00
* are also available at https://curl.se/docs/copyright.html.
*
* You may opt to use, copy, modify, merge, publish, distribute and/or sell
* copies of the Software, and permit persons to whom the Software is
* furnished to do so, under the terms of the COPYING file.
*
* This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
* KIND, either express or implied.
*
* SPDX-License-Identifier: curl
*
***************************************************************************/
#include <curl/curl.h>
#include "timeval.h"
struct connectdata;
cpool: rename "connection cache/conncache" to "Connection Pools/cpool" This is a better match for what they do and the general "cpool" var/function prefix works well. The pool now handles very long hostnames correctly. The following changes have been made: * 'struct connectdata', e.g. connections, keep new members named `destination` and ' destination_len' that fully specifies interface+port+hostname of where the connection is going to. This is used in the pool for "bundling" of connections with the same destination. There is no limit on the length any more. * Locking: all locks are done inside conncache.c when calling into the pool and released on return. This eliminates hazards of the callers keeping track. * 'struct connectbundle' is now internal to the pool. It is no longer referenced by a connection. * 'bundle->multiuse' no longer exists. HTTP/2 and 3 and TLS filters no longer need to set it. Instead, the multi checks on leaving MSTATE_CONNECT or MSTATE_CONNECTING if the connection is now multiplexed and new, e.g. not conn->bits.reuse. In that case the processing of pending handles is triggered. * The pool's init is provided with a callback to invoke on all connections being discarded. This allows the cleanups in `Curl_disconnect` to run, wherever it is decided to retire a connection. * Several pool operations can now be fully done with one call. Pruning dead connections, upkeep and checks on pool limits can now directly discard connections and need no longer return those to the caller for doing that (as we have now the callback described above). * Finding a connection for reuse is now done via `Curl_cpool_find()` and the caller provides callbacks to evaluate the connection candidates. * The 'Curl_cpool_check_limits()' now directly uses the max values that may be set in the transfer's multi. No need to pass them around. Curl_multi_max_host_connections() and Curl_multi_max_total_connections() are gone. * Add method 'Curl_node_llist()' to get the llist a node is in. Used in cpool to verify connection are indeed in the list (or not in any list) as they need to. I left the conncache.[ch] as is for now and also did not touch the documentation. If we update that outside the feature window, we can do this in a separate PR. Multi-thread safety is not achieved by this PR, but since more details on how pools operate are now "internal" it is a better starting point to go for this in the future. Closes #14662
2024-08-23 19:58:41 +08:00
struct Curl_easy;
lib: graceful connection shutdown When libcurl discards a connection there are two phases this may go through: "shutdown" and "closing". If a connection is aborted, the shutdown phase is skipped and it is closed right away. The connection filters attached to the connection implement the phases in their `do_shutdown()` and `do_close()` callbacks. Filters carry now a `shutdown` flags next to `connected` to keep track of the shutdown operation. Filters are shut down from top to bottom. If a filter is not connected, its shutdown is skipped. Notable filters that *do* something during shutdown are HTTP/2 and TLS. HTTP/2 sends the GOAWAY frame. TLS sends its close notify and expects to receive a close notify from the server. As sends and receives may EAGAIN on the network, a shutdown is often not successful right away and needs to poll the connection's socket(s). To facilitate this, such connections are placed on a new shutdown list inside the connection cache. Since managing this list requires the cooperation of a multi handle, only the connection cache belonging to a multi handle is used. If a connection was in another cache when being discarded, it is removed there and added to the multi's cache. If no multi handle is available at that time, the connection is shutdown and closed in a one-time, best-effort attempt. When a multi handle is destroyed, all connection still on the shutdown list are discarded with a final shutdown attempt and close. In curl debug builds, the environment variable `CURL_GRACEFUL_SHUTDOWN` can be set to make this graceful with a timeout in milliseconds given by the variable. The shutdown list is limited to the max number of connections configured for a multi cache. Set via CURLMOPT_MAX_TOTAL_CONNECTIONS. When the limit is reached, the oldest connection on the shutdown list is discarded. - In multi_wait() and multi_waitfds(), collect all connection caches involved (each transfer might carry its own) into a temporary list. Let each connection cache on the list contribute sockets and POLLIN/OUT events it's connections are waiting for. - in multi_perform() collect the connection caches the same way and let them peform their maintenance. This will make another non-blocking attempt to shutdown all connections on its shutdown list. - for event based multis (multi->socket_cb set), add the sockets and their poll events via the callback. When `multi_socket()` is invoked for a socket not known by an active transfer, forward this to the multi's cache for processing. On closing a connection, remove its socket(s) via the callback. TLS connection filters MUST NOT send close nofity messages in their `do_close()` implementation. The reason is that a TLS close notify signals a success. When a connection is aborted and skips its shutdown phase, the server needs to see a missing close notify to detect something has gone wrong. A graceful shutdown of FTP's data connection is performed implicitly before regarding the upload/download as complete and continuing on the control connection. For FTP without TLS, there is just the socket close happening. But with TLS, the sent/received close notify signals that the transfer is complete and healthy. Servers like `vsftpd` verify that and reject uploads without a TLS close notify. - added test_19_* for shutdown related tests - test_19_01 and test_19_02 test for TCP RST packets which happen without a graceful shutdown and should no longer appear otherwise. - add test_19_03 for handling shutdowns by the server - add test_19_04 for handling shutdowns by curl - add test_19_05 for event based shutdowny by server - add test_30_06/07 and test_31_06/07 for shutdown checks on FTP up- and downloads. Closes #13976
2024-06-19 18:40:06 +08:00
struct curl_pollfds;
struct curl_waitfds;
struct Curl_multi;
cpool: rename "connection cache/conncache" to "Connection Pools/cpool" This is a better match for what they do and the general "cpool" var/function prefix works well. The pool now handles very long hostnames correctly. The following changes have been made: * 'struct connectdata', e.g. connections, keep new members named `destination` and ' destination_len' that fully specifies interface+port+hostname of where the connection is going to. This is used in the pool for "bundling" of connections with the same destination. There is no limit on the length any more. * Locking: all locks are done inside conncache.c when calling into the pool and released on return. This eliminates hazards of the callers keeping track. * 'struct connectbundle' is now internal to the pool. It is no longer referenced by a connection. * 'bundle->multiuse' no longer exists. HTTP/2 and 3 and TLS filters no longer need to set it. Instead, the multi checks on leaving MSTATE_CONNECT or MSTATE_CONNECTING if the connection is now multiplexed and new, e.g. not conn->bits.reuse. In that case the processing of pending handles is triggered. * The pool's init is provided with a callback to invoke on all connections being discarded. This allows the cleanups in `Curl_disconnect` to run, wherever it is decided to retire a connection. * Several pool operations can now be fully done with one call. Pruning dead connections, upkeep and checks on pool limits can now directly discard connections and need no longer return those to the caller for doing that (as we have now the callback described above). * Finding a connection for reuse is now done via `Curl_cpool_find()` and the caller provides callbacks to evaluate the connection candidates. * The 'Curl_cpool_check_limits()' now directly uses the max values that may be set in the transfer's multi. No need to pass them around. Curl_multi_max_host_connections() and Curl_multi_max_total_connections() are gone. * Add method 'Curl_node_llist()' to get the llist a node is in. Used in cpool to verify connection are indeed in the list (or not in any list) as they need to. I left the conncache.[ch] as is for now and also did not touch the documentation. If we update that outside the feature window, we can do this in a separate PR. Multi-thread safety is not achieved by this PR, but since more details on how pools operate are now "internal" it is a better starting point to go for this in the future. Closes #14662
2024-08-23 19:58:41 +08:00
struct Curl_share;
lib: graceful connection shutdown When libcurl discards a connection there are two phases this may go through: "shutdown" and "closing". If a connection is aborted, the shutdown phase is skipped and it is closed right away. The connection filters attached to the connection implement the phases in their `do_shutdown()` and `do_close()` callbacks. Filters carry now a `shutdown` flags next to `connected` to keep track of the shutdown operation. Filters are shut down from top to bottom. If a filter is not connected, its shutdown is skipped. Notable filters that *do* something during shutdown are HTTP/2 and TLS. HTTP/2 sends the GOAWAY frame. TLS sends its close notify and expects to receive a close notify from the server. As sends and receives may EAGAIN on the network, a shutdown is often not successful right away and needs to poll the connection's socket(s). To facilitate this, such connections are placed on a new shutdown list inside the connection cache. Since managing this list requires the cooperation of a multi handle, only the connection cache belonging to a multi handle is used. If a connection was in another cache when being discarded, it is removed there and added to the multi's cache. If no multi handle is available at that time, the connection is shutdown and closed in a one-time, best-effort attempt. When a multi handle is destroyed, all connection still on the shutdown list are discarded with a final shutdown attempt and close. In curl debug builds, the environment variable `CURL_GRACEFUL_SHUTDOWN` can be set to make this graceful with a timeout in milliseconds given by the variable. The shutdown list is limited to the max number of connections configured for a multi cache. Set via CURLMOPT_MAX_TOTAL_CONNECTIONS. When the limit is reached, the oldest connection on the shutdown list is discarded. - In multi_wait() and multi_waitfds(), collect all connection caches involved (each transfer might carry its own) into a temporary list. Let each connection cache on the list contribute sockets and POLLIN/OUT events it's connections are waiting for. - in multi_perform() collect the connection caches the same way and let them peform their maintenance. This will make another non-blocking attempt to shutdown all connections on its shutdown list. - for event based multis (multi->socket_cb set), add the sockets and their poll events via the callback. When `multi_socket()` is invoked for a socket not known by an active transfer, forward this to the multi's cache for processing. On closing a connection, remove its socket(s) via the callback. TLS connection filters MUST NOT send close nofity messages in their `do_close()` implementation. The reason is that a TLS close notify signals a success. When a connection is aborted and skips its shutdown phase, the server needs to see a missing close notify to detect something has gone wrong. A graceful shutdown of FTP's data connection is performed implicitly before regarding the upload/download as complete and continuing on the control connection. For FTP without TLS, there is just the socket close happening. But with TLS, the sent/received close notify signals that the transfer is complete and healthy. Servers like `vsftpd` verify that and reject uploads without a TLS close notify. - added test_19_* for shutdown related tests - test_19_01 and test_19_02 test for TCP RST packets which happen without a graceful shutdown and should no longer appear otherwise. - add test_19_03 for handling shutdowns by the server - add test_19_04 for handling shutdowns by curl - add test_19_05 for event based shutdowny by server - add test_30_06/07 and test_31_06/07 for shutdown checks on FTP up- and downloads. Closes #13976
2024-06-19 18:40:06 +08:00
cpool: rename "connection cache/conncache" to "Connection Pools/cpool" This is a better match for what they do and the general "cpool" var/function prefix works well. The pool now handles very long hostnames correctly. The following changes have been made: * 'struct connectdata', e.g. connections, keep new members named `destination` and ' destination_len' that fully specifies interface+port+hostname of where the connection is going to. This is used in the pool for "bundling" of connections with the same destination. There is no limit on the length any more. * Locking: all locks are done inside conncache.c when calling into the pool and released on return. This eliminates hazards of the callers keeping track. * 'struct connectbundle' is now internal to the pool. It is no longer referenced by a connection. * 'bundle->multiuse' no longer exists. HTTP/2 and 3 and TLS filters no longer need to set it. Instead, the multi checks on leaving MSTATE_CONNECT or MSTATE_CONNECTING if the connection is now multiplexed and new, e.g. not conn->bits.reuse. In that case the processing of pending handles is triggered. * The pool's init is provided with a callback to invoke on all connections being discarded. This allows the cleanups in `Curl_disconnect` to run, wherever it is decided to retire a connection. * Several pool operations can now be fully done with one call. Pruning dead connections, upkeep and checks on pool limits can now directly discard connections and need no longer return those to the caller for doing that (as we have now the callback described above). * Finding a connection for reuse is now done via `Curl_cpool_find()` and the caller provides callbacks to evaluate the connection candidates. * The 'Curl_cpool_check_limits()' now directly uses the max values that may be set in the transfer's multi. No need to pass them around. Curl_multi_max_host_connections() and Curl_multi_max_total_connections() are gone. * Add method 'Curl_node_llist()' to get the llist a node is in. Used in cpool to verify connection are indeed in the list (or not in any list) as they need to. I left the conncache.[ch] as is for now and also did not touch the documentation. If we update that outside the feature window, we can do this in a separate PR. Multi-thread safety is not achieved by this PR, but since more details on how pools operate are now "internal" it is a better starting point to go for this in the future. Closes #14662
2024-08-23 19:58:41 +08:00
/**
* Callback invoked when disconnecting connections.
* @param data transfer last handling the connection, not attached
* @param conn the connection to discard
* @param aborted if the connection is being aborted
* @return if the connection is being aborted, e.g. should NOT perform
* a shutdown and just close.
**/
typedef bool Curl_cpool_disconnect_cb(struct Curl_easy *data,
struct connectdata *conn,
bool aborted);
struct cpool {
/* the pooled connections, bundled per destination */
struct Curl_hash dest2bundle;
size_t num_conn;
curl_off_t next_connection_id;
curl_off_t next_easy_id;
struct curltime last_cleanup;
cpool: rename "connection cache/conncache" to "Connection Pools/cpool" This is a better match for what they do and the general "cpool" var/function prefix works well. The pool now handles very long hostnames correctly. The following changes have been made: * 'struct connectdata', e.g. connections, keep new members named `destination` and ' destination_len' that fully specifies interface+port+hostname of where the connection is going to. This is used in the pool for "bundling" of connections with the same destination. There is no limit on the length any more. * Locking: all locks are done inside conncache.c when calling into the pool and released on return. This eliminates hazards of the callers keeping track. * 'struct connectbundle' is now internal to the pool. It is no longer referenced by a connection. * 'bundle->multiuse' no longer exists. HTTP/2 and 3 and TLS filters no longer need to set it. Instead, the multi checks on leaving MSTATE_CONNECT or MSTATE_CONNECTING if the connection is now multiplexed and new, e.g. not conn->bits.reuse. In that case the processing of pending handles is triggered. * The pool's init is provided with a callback to invoke on all connections being discarded. This allows the cleanups in `Curl_disconnect` to run, wherever it is decided to retire a connection. * Several pool operations can now be fully done with one call. Pruning dead connections, upkeep and checks on pool limits can now directly discard connections and need no longer return those to the caller for doing that (as we have now the callback described above). * Finding a connection for reuse is now done via `Curl_cpool_find()` and the caller provides callbacks to evaluate the connection candidates. * The 'Curl_cpool_check_limits()' now directly uses the max values that may be set in the transfer's multi. No need to pass them around. Curl_multi_max_host_connections() and Curl_multi_max_total_connections() are gone. * Add method 'Curl_node_llist()' to get the llist a node is in. Used in cpool to verify connection are indeed in the list (or not in any list) as they need to. I left the conncache.[ch] as is for now and also did not touch the documentation. If we update that outside the feature window, we can do this in a separate PR. Multi-thread safety is not achieved by this PR, but since more details on how pools operate are now "internal" it is a better starting point to go for this in the future. Closes #14662
2024-08-23 19:58:41 +08:00
struct Curl_llist shutdowns; /* The connections being shut down */
struct Curl_easy *idata; /* internal handle used for discard */
struct Curl_multi *multi; /* != NULL iff pool belongs to multi */
struct Curl_share *share; /* != NULL iff pool belongs to share */
Curl_cpool_disconnect_cb *disconnect_cb;
BIT(locked);
};
cpool: rename "connection cache/conncache" to "Connection Pools/cpool" This is a better match for what they do and the general "cpool" var/function prefix works well. The pool now handles very long hostnames correctly. The following changes have been made: * 'struct connectdata', e.g. connections, keep new members named `destination` and ' destination_len' that fully specifies interface+port+hostname of where the connection is going to. This is used in the pool for "bundling" of connections with the same destination. There is no limit on the length any more. * Locking: all locks are done inside conncache.c when calling into the pool and released on return. This eliminates hazards of the callers keeping track. * 'struct connectbundle' is now internal to the pool. It is no longer referenced by a connection. * 'bundle->multiuse' no longer exists. HTTP/2 and 3 and TLS filters no longer need to set it. Instead, the multi checks on leaving MSTATE_CONNECT or MSTATE_CONNECTING if the connection is now multiplexed and new, e.g. not conn->bits.reuse. In that case the processing of pending handles is triggered. * The pool's init is provided with a callback to invoke on all connections being discarded. This allows the cleanups in `Curl_disconnect` to run, wherever it is decided to retire a connection. * Several pool operations can now be fully done with one call. Pruning dead connections, upkeep and checks on pool limits can now directly discard connections and need no longer return those to the caller for doing that (as we have now the callback described above). * Finding a connection for reuse is now done via `Curl_cpool_find()` and the caller provides callbacks to evaluate the connection candidates. * The 'Curl_cpool_check_limits()' now directly uses the max values that may be set in the transfer's multi. No need to pass them around. Curl_multi_max_host_connections() and Curl_multi_max_total_connections() are gone. * Add method 'Curl_node_llist()' to get the llist a node is in. Used in cpool to verify connection are indeed in the list (or not in any list) as they need to. I left the conncache.[ch] as is for now and also did not touch the documentation. If we update that outside the feature window, we can do this in a separate PR. Multi-thread safety is not achieved by this PR, but since more details on how pools operate are now "internal" it is a better starting point to go for this in the future. Closes #14662
2024-08-23 19:58:41 +08:00
/* Init the pool, pass multi only if pool is owned by it.
lib: graceful connection shutdown When libcurl discards a connection there are two phases this may go through: "shutdown" and "closing". If a connection is aborted, the shutdown phase is skipped and it is closed right away. The connection filters attached to the connection implement the phases in their `do_shutdown()` and `do_close()` callbacks. Filters carry now a `shutdown` flags next to `connected` to keep track of the shutdown operation. Filters are shut down from top to bottom. If a filter is not connected, its shutdown is skipped. Notable filters that *do* something during shutdown are HTTP/2 and TLS. HTTP/2 sends the GOAWAY frame. TLS sends its close notify and expects to receive a close notify from the server. As sends and receives may EAGAIN on the network, a shutdown is often not successful right away and needs to poll the connection's socket(s). To facilitate this, such connections are placed on a new shutdown list inside the connection cache. Since managing this list requires the cooperation of a multi handle, only the connection cache belonging to a multi handle is used. If a connection was in another cache when being discarded, it is removed there and added to the multi's cache. If no multi handle is available at that time, the connection is shutdown and closed in a one-time, best-effort attempt. When a multi handle is destroyed, all connection still on the shutdown list are discarded with a final shutdown attempt and close. In curl debug builds, the environment variable `CURL_GRACEFUL_SHUTDOWN` can be set to make this graceful with a timeout in milliseconds given by the variable. The shutdown list is limited to the max number of connections configured for a multi cache. Set via CURLMOPT_MAX_TOTAL_CONNECTIONS. When the limit is reached, the oldest connection on the shutdown list is discarded. - In multi_wait() and multi_waitfds(), collect all connection caches involved (each transfer might carry its own) into a temporary list. Let each connection cache on the list contribute sockets and POLLIN/OUT events it's connections are waiting for. - in multi_perform() collect the connection caches the same way and let them peform their maintenance. This will make another non-blocking attempt to shutdown all connections on its shutdown list. - for event based multis (multi->socket_cb set), add the sockets and their poll events via the callback. When `multi_socket()` is invoked for a socket not known by an active transfer, forward this to the multi's cache for processing. On closing a connection, remove its socket(s) via the callback. TLS connection filters MUST NOT send close nofity messages in their `do_close()` implementation. The reason is that a TLS close notify signals a success. When a connection is aborted and skips its shutdown phase, the server needs to see a missing close notify to detect something has gone wrong. A graceful shutdown of FTP's data connection is performed implicitly before regarding the upload/download as complete and continuing on the control connection. For FTP without TLS, there is just the socket close happening. But with TLS, the sent/received close notify signals that the transfer is complete and healthy. Servers like `vsftpd` verify that and reject uploads without a TLS close notify. - added test_19_* for shutdown related tests - test_19_01 and test_19_02 test for TCP RST packets which happen without a graceful shutdown and should no longer appear otherwise. - add test_19_03 for handling shutdowns by the server - add test_19_04 for handling shutdowns by curl - add test_19_05 for event based shutdowny by server - add test_30_06/07 and test_31_06/07 for shutdown checks on FTP up- and downloads. Closes #13976
2024-06-19 18:40:06 +08:00
* returns 1 on error, 0 is fine.
*/
cpool: rename "connection cache/conncache" to "Connection Pools/cpool" This is a better match for what they do and the general "cpool" var/function prefix works well. The pool now handles very long hostnames correctly. The following changes have been made: * 'struct connectdata', e.g. connections, keep new members named `destination` and ' destination_len' that fully specifies interface+port+hostname of where the connection is going to. This is used in the pool for "bundling" of connections with the same destination. There is no limit on the length any more. * Locking: all locks are done inside conncache.c when calling into the pool and released on return. This eliminates hazards of the callers keeping track. * 'struct connectbundle' is now internal to the pool. It is no longer referenced by a connection. * 'bundle->multiuse' no longer exists. HTTP/2 and 3 and TLS filters no longer need to set it. Instead, the multi checks on leaving MSTATE_CONNECT or MSTATE_CONNECTING if the connection is now multiplexed and new, e.g. not conn->bits.reuse. In that case the processing of pending handles is triggered. * The pool's init is provided with a callback to invoke on all connections being discarded. This allows the cleanups in `Curl_disconnect` to run, wherever it is decided to retire a connection. * Several pool operations can now be fully done with one call. Pruning dead connections, upkeep and checks on pool limits can now directly discard connections and need no longer return those to the caller for doing that (as we have now the callback described above). * Finding a connection for reuse is now done via `Curl_cpool_find()` and the caller provides callbacks to evaluate the connection candidates. * The 'Curl_cpool_check_limits()' now directly uses the max values that may be set in the transfer's multi. No need to pass them around. Curl_multi_max_host_connections() and Curl_multi_max_total_connections() are gone. * Add method 'Curl_node_llist()' to get the llist a node is in. Used in cpool to verify connection are indeed in the list (or not in any list) as they need to. I left the conncache.[ch] as is for now and also did not touch the documentation. If we update that outside the feature window, we can do this in a separate PR. Multi-thread safety is not achieved by this PR, but since more details on how pools operate are now "internal" it is a better starting point to go for this in the future. Closes #14662
2024-08-23 19:58:41 +08:00
int Curl_cpool_init(struct cpool *cpool,
Curl_cpool_disconnect_cb *disconnect_cb,
struct Curl_multi *multi,
struct Curl_share *share,
size_t size);
/* Destroy all connections and free all members */
void Curl_cpool_destroy(struct cpool *connc);
/* Init the transfer to be used within its connection pool.
* Assigns `data->id`. */
void Curl_cpool_xfer_init(struct Curl_easy *data);
lib: graceful connection shutdown When libcurl discards a connection there are two phases this may go through: "shutdown" and "closing". If a connection is aborted, the shutdown phase is skipped and it is closed right away. The connection filters attached to the connection implement the phases in their `do_shutdown()` and `do_close()` callbacks. Filters carry now a `shutdown` flags next to `connected` to keep track of the shutdown operation. Filters are shut down from top to bottom. If a filter is not connected, its shutdown is skipped. Notable filters that *do* something during shutdown are HTTP/2 and TLS. HTTP/2 sends the GOAWAY frame. TLS sends its close notify and expects to receive a close notify from the server. As sends and receives may EAGAIN on the network, a shutdown is often not successful right away and needs to poll the connection's socket(s). To facilitate this, such connections are placed on a new shutdown list inside the connection cache. Since managing this list requires the cooperation of a multi handle, only the connection cache belonging to a multi handle is used. If a connection was in another cache when being discarded, it is removed there and added to the multi's cache. If no multi handle is available at that time, the connection is shutdown and closed in a one-time, best-effort attempt. When a multi handle is destroyed, all connection still on the shutdown list are discarded with a final shutdown attempt and close. In curl debug builds, the environment variable `CURL_GRACEFUL_SHUTDOWN` can be set to make this graceful with a timeout in milliseconds given by the variable. The shutdown list is limited to the max number of connections configured for a multi cache. Set via CURLMOPT_MAX_TOTAL_CONNECTIONS. When the limit is reached, the oldest connection on the shutdown list is discarded. - In multi_wait() and multi_waitfds(), collect all connection caches involved (each transfer might carry its own) into a temporary list. Let each connection cache on the list contribute sockets and POLLIN/OUT events it's connections are waiting for. - in multi_perform() collect the connection caches the same way and let them peform their maintenance. This will make another non-blocking attempt to shutdown all connections on its shutdown list. - for event based multis (multi->socket_cb set), add the sockets and their poll events via the callback. When `multi_socket()` is invoked for a socket not known by an active transfer, forward this to the multi's cache for processing. On closing a connection, remove its socket(s) via the callback. TLS connection filters MUST NOT send close nofity messages in their `do_close()` implementation. The reason is that a TLS close notify signals a success. When a connection is aborted and skips its shutdown phase, the server needs to see a missing close notify to detect something has gone wrong. A graceful shutdown of FTP's data connection is performed implicitly before regarding the upload/download as complete and continuing on the control connection. For FTP without TLS, there is just the socket close happening. But with TLS, the sent/received close notify signals that the transfer is complete and healthy. Servers like `vsftpd` verify that and reject uploads without a TLS close notify. - added test_19_* for shutdown related tests - test_19_01 and test_19_02 test for TCP RST packets which happen without a graceful shutdown and should no longer appear otherwise. - add test_19_03 for handling shutdowns by the server - add test_19_04 for handling shutdowns by curl - add test_19_05 for event based shutdowny by server - add test_30_06/07 and test_31_06/07 for shutdown checks on FTP up- and downloads. Closes #13976
2024-06-19 18:40:06 +08:00
/**
cpool: rename "connection cache/conncache" to "Connection Pools/cpool" This is a better match for what they do and the general "cpool" var/function prefix works well. The pool now handles very long hostnames correctly. The following changes have been made: * 'struct connectdata', e.g. connections, keep new members named `destination` and ' destination_len' that fully specifies interface+port+hostname of where the connection is going to. This is used in the pool for "bundling" of connections with the same destination. There is no limit on the length any more. * Locking: all locks are done inside conncache.c when calling into the pool and released on return. This eliminates hazards of the callers keeping track. * 'struct connectbundle' is now internal to the pool. It is no longer referenced by a connection. * 'bundle->multiuse' no longer exists. HTTP/2 and 3 and TLS filters no longer need to set it. Instead, the multi checks on leaving MSTATE_CONNECT or MSTATE_CONNECTING if the connection is now multiplexed and new, e.g. not conn->bits.reuse. In that case the processing of pending handles is triggered. * The pool's init is provided with a callback to invoke on all connections being discarded. This allows the cleanups in `Curl_disconnect` to run, wherever it is decided to retire a connection. * Several pool operations can now be fully done with one call. Pruning dead connections, upkeep and checks on pool limits can now directly discard connections and need no longer return those to the caller for doing that (as we have now the callback described above). * Finding a connection for reuse is now done via `Curl_cpool_find()` and the caller provides callbacks to evaluate the connection candidates. * The 'Curl_cpool_check_limits()' now directly uses the max values that may be set in the transfer's multi. No need to pass them around. Curl_multi_max_host_connections() and Curl_multi_max_total_connections() are gone. * Add method 'Curl_node_llist()' to get the llist a node is in. Used in cpool to verify connection are indeed in the list (or not in any list) as they need to. I left the conncache.[ch] as is for now and also did not touch the documentation. If we update that outside the feature window, we can do this in a separate PR. Multi-thread safety is not achieved by this PR, but since more details on how pools operate are now "internal" it is a better starting point to go for this in the future. Closes #14662
2024-08-23 19:58:41 +08:00
* Get the connection with the given id from the transfer's pool.
lib: graceful connection shutdown When libcurl discards a connection there are two phases this may go through: "shutdown" and "closing". If a connection is aborted, the shutdown phase is skipped and it is closed right away. The connection filters attached to the connection implement the phases in their `do_shutdown()` and `do_close()` callbacks. Filters carry now a `shutdown` flags next to `connected` to keep track of the shutdown operation. Filters are shut down from top to bottom. If a filter is not connected, its shutdown is skipped. Notable filters that *do* something during shutdown are HTTP/2 and TLS. HTTP/2 sends the GOAWAY frame. TLS sends its close notify and expects to receive a close notify from the server. As sends and receives may EAGAIN on the network, a shutdown is often not successful right away and needs to poll the connection's socket(s). To facilitate this, such connections are placed on a new shutdown list inside the connection cache. Since managing this list requires the cooperation of a multi handle, only the connection cache belonging to a multi handle is used. If a connection was in another cache when being discarded, it is removed there and added to the multi's cache. If no multi handle is available at that time, the connection is shutdown and closed in a one-time, best-effort attempt. When a multi handle is destroyed, all connection still on the shutdown list are discarded with a final shutdown attempt and close. In curl debug builds, the environment variable `CURL_GRACEFUL_SHUTDOWN` can be set to make this graceful with a timeout in milliseconds given by the variable. The shutdown list is limited to the max number of connections configured for a multi cache. Set via CURLMOPT_MAX_TOTAL_CONNECTIONS. When the limit is reached, the oldest connection on the shutdown list is discarded. - In multi_wait() and multi_waitfds(), collect all connection caches involved (each transfer might carry its own) into a temporary list. Let each connection cache on the list contribute sockets and POLLIN/OUT events it's connections are waiting for. - in multi_perform() collect the connection caches the same way and let them peform their maintenance. This will make another non-blocking attempt to shutdown all connections on its shutdown list. - for event based multis (multi->socket_cb set), add the sockets and their poll events via the callback. When `multi_socket()` is invoked for a socket not known by an active transfer, forward this to the multi's cache for processing. On closing a connection, remove its socket(s) via the callback. TLS connection filters MUST NOT send close nofity messages in their `do_close()` implementation. The reason is that a TLS close notify signals a success. When a connection is aborted and skips its shutdown phase, the server needs to see a missing close notify to detect something has gone wrong. A graceful shutdown of FTP's data connection is performed implicitly before regarding the upload/download as complete and continuing on the control connection. For FTP without TLS, there is just the socket close happening. But with TLS, the sent/received close notify signals that the transfer is complete and healthy. Servers like `vsftpd` verify that and reject uploads without a TLS close notify. - added test_19_* for shutdown related tests - test_19_01 and test_19_02 test for TCP RST packets which happen without a graceful shutdown and should no longer appear otherwise. - add test_19_03 for handling shutdowns by the server - add test_19_04 for handling shutdowns by curl - add test_19_05 for event based shutdowny by server - add test_30_06/07 and test_31_06/07 for shutdown checks on FTP up- and downloads. Closes #13976
2024-06-19 18:40:06 +08:00
*/
cpool: rename "connection cache/conncache" to "Connection Pools/cpool" This is a better match for what they do and the general "cpool" var/function prefix works well. The pool now handles very long hostnames correctly. The following changes have been made: * 'struct connectdata', e.g. connections, keep new members named `destination` and ' destination_len' that fully specifies interface+port+hostname of where the connection is going to. This is used in the pool for "bundling" of connections with the same destination. There is no limit on the length any more. * Locking: all locks are done inside conncache.c when calling into the pool and released on return. This eliminates hazards of the callers keeping track. * 'struct connectbundle' is now internal to the pool. It is no longer referenced by a connection. * 'bundle->multiuse' no longer exists. HTTP/2 and 3 and TLS filters no longer need to set it. Instead, the multi checks on leaving MSTATE_CONNECT or MSTATE_CONNECTING if the connection is now multiplexed and new, e.g. not conn->bits.reuse. In that case the processing of pending handles is triggered. * The pool's init is provided with a callback to invoke on all connections being discarded. This allows the cleanups in `Curl_disconnect` to run, wherever it is decided to retire a connection. * Several pool operations can now be fully done with one call. Pruning dead connections, upkeep and checks on pool limits can now directly discard connections and need no longer return those to the caller for doing that (as we have now the callback described above). * Finding a connection for reuse is now done via `Curl_cpool_find()` and the caller provides callbacks to evaluate the connection candidates. * The 'Curl_cpool_check_limits()' now directly uses the max values that may be set in the transfer's multi. No need to pass them around. Curl_multi_max_host_connections() and Curl_multi_max_total_connections() are gone. * Add method 'Curl_node_llist()' to get the llist a node is in. Used in cpool to verify connection are indeed in the list (or not in any list) as they need to. I left the conncache.[ch] as is for now and also did not touch the documentation. If we update that outside the feature window, we can do this in a separate PR. Multi-thread safety is not achieved by this PR, but since more details on how pools operate are now "internal" it is a better starting point to go for this in the future. Closes #14662
2024-08-23 19:58:41 +08:00
struct connectdata *Curl_cpool_get_conn(struct Curl_easy *data,
curl_off_t conn_id);
CURLcode Curl_cpool_add_conn(struct Curl_easy *data,
struct connectdata *conn) WARN_UNUSED_RESULT;
lib: graceful connection shutdown When libcurl discards a connection there are two phases this may go through: "shutdown" and "closing". If a connection is aborted, the shutdown phase is skipped and it is closed right away. The connection filters attached to the connection implement the phases in their `do_shutdown()` and `do_close()` callbacks. Filters carry now a `shutdown` flags next to `connected` to keep track of the shutdown operation. Filters are shut down from top to bottom. If a filter is not connected, its shutdown is skipped. Notable filters that *do* something during shutdown are HTTP/2 and TLS. HTTP/2 sends the GOAWAY frame. TLS sends its close notify and expects to receive a close notify from the server. As sends and receives may EAGAIN on the network, a shutdown is often not successful right away and needs to poll the connection's socket(s). To facilitate this, such connections are placed on a new shutdown list inside the connection cache. Since managing this list requires the cooperation of a multi handle, only the connection cache belonging to a multi handle is used. If a connection was in another cache when being discarded, it is removed there and added to the multi's cache. If no multi handle is available at that time, the connection is shutdown and closed in a one-time, best-effort attempt. When a multi handle is destroyed, all connection still on the shutdown list are discarded with a final shutdown attempt and close. In curl debug builds, the environment variable `CURL_GRACEFUL_SHUTDOWN` can be set to make this graceful with a timeout in milliseconds given by the variable. The shutdown list is limited to the max number of connections configured for a multi cache. Set via CURLMOPT_MAX_TOTAL_CONNECTIONS. When the limit is reached, the oldest connection on the shutdown list is discarded. - In multi_wait() and multi_waitfds(), collect all connection caches involved (each transfer might carry its own) into a temporary list. Let each connection cache on the list contribute sockets and POLLIN/OUT events it's connections are waiting for. - in multi_perform() collect the connection caches the same way and let them peform their maintenance. This will make another non-blocking attempt to shutdown all connections on its shutdown list. - for event based multis (multi->socket_cb set), add the sockets and their poll events via the callback. When `multi_socket()` is invoked for a socket not known by an active transfer, forward this to the multi's cache for processing. On closing a connection, remove its socket(s) via the callback. TLS connection filters MUST NOT send close nofity messages in their `do_close()` implementation. The reason is that a TLS close notify signals a success. When a connection is aborted and skips its shutdown phase, the server needs to see a missing close notify to detect something has gone wrong. A graceful shutdown of FTP's data connection is performed implicitly before regarding the upload/download as complete and continuing on the control connection. For FTP without TLS, there is just the socket close happening. But with TLS, the sent/received close notify signals that the transfer is complete and healthy. Servers like `vsftpd` verify that and reject uploads without a TLS close notify. - added test_19_* for shutdown related tests - test_19_01 and test_19_02 test for TCP RST packets which happen without a graceful shutdown and should no longer appear otherwise. - add test_19_03 for handling shutdowns by the server - add test_19_04 for handling shutdowns by curl - add test_19_05 for event based shutdowny by server - add test_30_06/07 and test_31_06/07 for shutdown checks on FTP up- and downloads. Closes #13976
2024-06-19 18:40:06 +08:00
/**
cpool: rename "connection cache/conncache" to "Connection Pools/cpool" This is a better match for what they do and the general "cpool" var/function prefix works well. The pool now handles very long hostnames correctly. The following changes have been made: * 'struct connectdata', e.g. connections, keep new members named `destination` and ' destination_len' that fully specifies interface+port+hostname of where the connection is going to. This is used in the pool for "bundling" of connections with the same destination. There is no limit on the length any more. * Locking: all locks are done inside conncache.c when calling into the pool and released on return. This eliminates hazards of the callers keeping track. * 'struct connectbundle' is now internal to the pool. It is no longer referenced by a connection. * 'bundle->multiuse' no longer exists. HTTP/2 and 3 and TLS filters no longer need to set it. Instead, the multi checks on leaving MSTATE_CONNECT or MSTATE_CONNECTING if the connection is now multiplexed and new, e.g. not conn->bits.reuse. In that case the processing of pending handles is triggered. * The pool's init is provided with a callback to invoke on all connections being discarded. This allows the cleanups in `Curl_disconnect` to run, wherever it is decided to retire a connection. * Several pool operations can now be fully done with one call. Pruning dead connections, upkeep and checks on pool limits can now directly discard connections and need no longer return those to the caller for doing that (as we have now the callback described above). * Finding a connection for reuse is now done via `Curl_cpool_find()` and the caller provides callbacks to evaluate the connection candidates. * The 'Curl_cpool_check_limits()' now directly uses the max values that may be set in the transfer's multi. No need to pass them around. Curl_multi_max_host_connections() and Curl_multi_max_total_connections() are gone. * Add method 'Curl_node_llist()' to get the llist a node is in. Used in cpool to verify connection are indeed in the list (or not in any list) as they need to. I left the conncache.[ch] as is for now and also did not touch the documentation. If we update that outside the feature window, we can do this in a separate PR. Multi-thread safety is not achieved by this PR, but since more details on how pools operate are now "internal" it is a better starting point to go for this in the future. Closes #14662
2024-08-23 19:58:41 +08:00
* Return if the pool has reached its configured limits for adding
* the given connection. Will try to discard the oldest, idle
* connections to make space.
lib: graceful connection shutdown When libcurl discards a connection there are two phases this may go through: "shutdown" and "closing". If a connection is aborted, the shutdown phase is skipped and it is closed right away. The connection filters attached to the connection implement the phases in their `do_shutdown()` and `do_close()` callbacks. Filters carry now a `shutdown` flags next to `connected` to keep track of the shutdown operation. Filters are shut down from top to bottom. If a filter is not connected, its shutdown is skipped. Notable filters that *do* something during shutdown are HTTP/2 and TLS. HTTP/2 sends the GOAWAY frame. TLS sends its close notify and expects to receive a close notify from the server. As sends and receives may EAGAIN on the network, a shutdown is often not successful right away and needs to poll the connection's socket(s). To facilitate this, such connections are placed on a new shutdown list inside the connection cache. Since managing this list requires the cooperation of a multi handle, only the connection cache belonging to a multi handle is used. If a connection was in another cache when being discarded, it is removed there and added to the multi's cache. If no multi handle is available at that time, the connection is shutdown and closed in a one-time, best-effort attempt. When a multi handle is destroyed, all connection still on the shutdown list are discarded with a final shutdown attempt and close. In curl debug builds, the environment variable `CURL_GRACEFUL_SHUTDOWN` can be set to make this graceful with a timeout in milliseconds given by the variable. The shutdown list is limited to the max number of connections configured for a multi cache. Set via CURLMOPT_MAX_TOTAL_CONNECTIONS. When the limit is reached, the oldest connection on the shutdown list is discarded. - In multi_wait() and multi_waitfds(), collect all connection caches involved (each transfer might carry its own) into a temporary list. Let each connection cache on the list contribute sockets and POLLIN/OUT events it's connections are waiting for. - in multi_perform() collect the connection caches the same way and let them peform their maintenance. This will make another non-blocking attempt to shutdown all connections on its shutdown list. - for event based multis (multi->socket_cb set), add the sockets and their poll events via the callback. When `multi_socket()` is invoked for a socket not known by an active transfer, forward this to the multi's cache for processing. On closing a connection, remove its socket(s) via the callback. TLS connection filters MUST NOT send close nofity messages in their `do_close()` implementation. The reason is that a TLS close notify signals a success. When a connection is aborted and skips its shutdown phase, the server needs to see a missing close notify to detect something has gone wrong. A graceful shutdown of FTP's data connection is performed implicitly before regarding the upload/download as complete and continuing on the control connection. For FTP without TLS, there is just the socket close happening. But with TLS, the sent/received close notify signals that the transfer is complete and healthy. Servers like `vsftpd` verify that and reject uploads without a TLS close notify. - added test_19_* for shutdown related tests - test_19_01 and test_19_02 test for TCP RST packets which happen without a graceful shutdown and should no longer appear otherwise. - add test_19_03 for handling shutdowns by the server - add test_19_04 for handling shutdowns by curl - add test_19_05 for event based shutdowny by server - add test_30_06/07 and test_31_06/07 for shutdown checks on FTP up- and downloads. Closes #13976
2024-06-19 18:40:06 +08:00
*/
cpool: rename "connection cache/conncache" to "Connection Pools/cpool" This is a better match for what they do and the general "cpool" var/function prefix works well. The pool now handles very long hostnames correctly. The following changes have been made: * 'struct connectdata', e.g. connections, keep new members named `destination` and ' destination_len' that fully specifies interface+port+hostname of where the connection is going to. This is used in the pool for "bundling" of connections with the same destination. There is no limit on the length any more. * Locking: all locks are done inside conncache.c when calling into the pool and released on return. This eliminates hazards of the callers keeping track. * 'struct connectbundle' is now internal to the pool. It is no longer referenced by a connection. * 'bundle->multiuse' no longer exists. HTTP/2 and 3 and TLS filters no longer need to set it. Instead, the multi checks on leaving MSTATE_CONNECT or MSTATE_CONNECTING if the connection is now multiplexed and new, e.g. not conn->bits.reuse. In that case the processing of pending handles is triggered. * The pool's init is provided with a callback to invoke on all connections being discarded. This allows the cleanups in `Curl_disconnect` to run, wherever it is decided to retire a connection. * Several pool operations can now be fully done with one call. Pruning dead connections, upkeep and checks on pool limits can now directly discard connections and need no longer return those to the caller for doing that (as we have now the callback described above). * Finding a connection for reuse is now done via `Curl_cpool_find()` and the caller provides callbacks to evaluate the connection candidates. * The 'Curl_cpool_check_limits()' now directly uses the max values that may be set in the transfer's multi. No need to pass them around. Curl_multi_max_host_connections() and Curl_multi_max_total_connections() are gone. * Add method 'Curl_node_llist()' to get the llist a node is in. Used in cpool to verify connection are indeed in the list (or not in any list) as they need to. I left the conncache.[ch] as is for now and also did not touch the documentation. If we update that outside the feature window, we can do this in a separate PR. Multi-thread safety is not achieved by this PR, but since more details on how pools operate are now "internal" it is a better starting point to go for this in the future. Closes #14662
2024-08-23 19:58:41 +08:00
#define CPOOL_LIMIT_OK 0
#define CPOOL_LIMIT_DEST 1
#define CPOOL_LIMIT_TOTAL 2
int Curl_cpool_check_limits(struct Curl_easy *data,
struct connectdata *conn);
/* Return of conn is suitable. If so, stops iteration. */
typedef bool Curl_cpool_conn_match_cb(struct connectdata *conn,
void *userdata);
/* Act on the result of the find, may override it. */
typedef bool Curl_cpool_done_match_cb(bool result, void *userdata);
/**
* Find a connection in the pool matching `destination`.
* All callbacks are invoked while the pool's lock is held.
* @param data current transfer
* @param destination match agaonst `conn->destination` in pool
* @param dest_len destination length, including terminating NUL
* @param conn_cb must be present, called for each connection in the
* bundle until it returns TRUE
* @param result_cb if not NULL, is called at the end with the result
* of the `conn_cb` or FALSE if never called.
* @return combined result of last conn_db and result_cb or FALSE if no
connections were present.
*/
bool Curl_cpool_find(struct Curl_easy *data,
const char *destination, size_t dest_len,
Curl_cpool_conn_match_cb *conn_cb,
Curl_cpool_done_match_cb *done_cb,
void *userdata);
/*
* A connection (already in the pool) is now idle. Do any
* cleanups in regard to the pool's limits.
*
* Return TRUE if idle connection kept in pool, FALSE if closed.
*/
bool Curl_cpool_conn_now_idle(struct Curl_easy *data,
struct connectdata *conn);
lib: graceful connection shutdown When libcurl discards a connection there are two phases this may go through: "shutdown" and "closing". If a connection is aborted, the shutdown phase is skipped and it is closed right away. The connection filters attached to the connection implement the phases in their `do_shutdown()` and `do_close()` callbacks. Filters carry now a `shutdown` flags next to `connected` to keep track of the shutdown operation. Filters are shut down from top to bottom. If a filter is not connected, its shutdown is skipped. Notable filters that *do* something during shutdown are HTTP/2 and TLS. HTTP/2 sends the GOAWAY frame. TLS sends its close notify and expects to receive a close notify from the server. As sends and receives may EAGAIN on the network, a shutdown is often not successful right away and needs to poll the connection's socket(s). To facilitate this, such connections are placed on a new shutdown list inside the connection cache. Since managing this list requires the cooperation of a multi handle, only the connection cache belonging to a multi handle is used. If a connection was in another cache when being discarded, it is removed there and added to the multi's cache. If no multi handle is available at that time, the connection is shutdown and closed in a one-time, best-effort attempt. When a multi handle is destroyed, all connection still on the shutdown list are discarded with a final shutdown attempt and close. In curl debug builds, the environment variable `CURL_GRACEFUL_SHUTDOWN` can be set to make this graceful with a timeout in milliseconds given by the variable. The shutdown list is limited to the max number of connections configured for a multi cache. Set via CURLMOPT_MAX_TOTAL_CONNECTIONS. When the limit is reached, the oldest connection on the shutdown list is discarded. - In multi_wait() and multi_waitfds(), collect all connection caches involved (each transfer might carry its own) into a temporary list. Let each connection cache on the list contribute sockets and POLLIN/OUT events it's connections are waiting for. - in multi_perform() collect the connection caches the same way and let them peform their maintenance. This will make another non-blocking attempt to shutdown all connections on its shutdown list. - for event based multis (multi->socket_cb set), add the sockets and their poll events via the callback. When `multi_socket()` is invoked for a socket not known by an active transfer, forward this to the multi's cache for processing. On closing a connection, remove its socket(s) via the callback. TLS connection filters MUST NOT send close nofity messages in their `do_close()` implementation. The reason is that a TLS close notify signals a success. When a connection is aborted and skips its shutdown phase, the server needs to see a missing close notify to detect something has gone wrong. A graceful shutdown of FTP's data connection is performed implicitly before regarding the upload/download as complete and continuing on the control connection. For FTP without TLS, there is just the socket close happening. But with TLS, the sent/received close notify signals that the transfer is complete and healthy. Servers like `vsftpd` verify that and reject uploads without a TLS close notify. - added test_19_* for shutdown related tests - test_19_01 and test_19_02 test for TCP RST packets which happen without a graceful shutdown and should no longer appear otherwise. - add test_19_03 for handling shutdowns by the server - add test_19_04 for handling shutdowns by curl - add test_19_05 for event based shutdowny by server - add test_30_06/07 and test_31_06/07 for shutdown checks on FTP up- and downloads. Closes #13976
2024-06-19 18:40:06 +08:00
/**
cpool: rename "connection cache/conncache" to "Connection Pools/cpool" This is a better match for what they do and the general "cpool" var/function prefix works well. The pool now handles very long hostnames correctly. The following changes have been made: * 'struct connectdata', e.g. connections, keep new members named `destination` and ' destination_len' that fully specifies interface+port+hostname of where the connection is going to. This is used in the pool for "bundling" of connections with the same destination. There is no limit on the length any more. * Locking: all locks are done inside conncache.c when calling into the pool and released on return. This eliminates hazards of the callers keeping track. * 'struct connectbundle' is now internal to the pool. It is no longer referenced by a connection. * 'bundle->multiuse' no longer exists. HTTP/2 and 3 and TLS filters no longer need to set it. Instead, the multi checks on leaving MSTATE_CONNECT or MSTATE_CONNECTING if the connection is now multiplexed and new, e.g. not conn->bits.reuse. In that case the processing of pending handles is triggered. * The pool's init is provided with a callback to invoke on all connections being discarded. This allows the cleanups in `Curl_disconnect` to run, wherever it is decided to retire a connection. * Several pool operations can now be fully done with one call. Pruning dead connections, upkeep and checks on pool limits can now directly discard connections and need no longer return those to the caller for doing that (as we have now the callback described above). * Finding a connection for reuse is now done via `Curl_cpool_find()` and the caller provides callbacks to evaluate the connection candidates. * The 'Curl_cpool_check_limits()' now directly uses the max values that may be set in the transfer's multi. No need to pass them around. Curl_multi_max_host_connections() and Curl_multi_max_total_connections() are gone. * Add method 'Curl_node_llist()' to get the llist a node is in. Used in cpool to verify connection are indeed in the list (or not in any list) as they need to. I left the conncache.[ch] as is for now and also did not touch the documentation. If we update that outside the feature window, we can do this in a separate PR. Multi-thread safety is not achieved by this PR, but since more details on how pools operate are now "internal" it is a better starting point to go for this in the future. Closes #14662
2024-08-23 19:58:41 +08:00
* Remove the connection from the pool and tear it down.
* If `aborted` is FALSE, the connection will be shut down first
* before closing and destroying it.
* If the shutdown is not immediately complete, the connection
* will be placed into the pool's shutdown queue.
*/
void Curl_cpool_disconnect(struct Curl_easy *data,
struct connectdata *conn,
bool aborted);
/**
* This function scans the data's connection pool for half-open/dead
* connections, closes and removes them.
* The cleanup is done at most once per second.
*
* When called, this transfer has no connection attached.
*/
void Curl_cpool_prune_dead(struct Curl_easy *data);
/**
* Perform upkeep actions on connections in the transfer's pool.
*/
CURLcode Curl_cpool_upkeep(void *data);
typedef void Curl_cpool_conn_do_cb(struct connectdata *conn,
struct Curl_easy *data,
void *cbdata);
/**
* Invoke the callback on the pool's connection with the
* given connection id (if it exists).
*/
void Curl_cpool_do_by_id(struct Curl_easy *data,
curl_off_t conn_id,
Curl_cpool_conn_do_cb *cb, void *cbdata);
/**
* Invoked the callback for the given data + connection under the
* connection pool's lock.
* The callback is always invoked, even if the transfer has no connection
* pool associated.
*/
void Curl_cpool_do_locked(struct Curl_easy *data,
struct connectdata *conn,
Curl_cpool_conn_do_cb *cb, void *cbdata);
/**
* Add sockets and POLLIN/OUT flags for connections handled by the pool.
*/
CURLcode Curl_cpool_add_pollfds(struct cpool *connc,
struct curl_pollfds *cpfds);
CURLcode Curl_cpool_add_waitfds(struct cpool *connc,
struct curl_waitfds *cwfds);
/**
* Perform maintenance on connections in the pool. Specifically,
lib: graceful connection shutdown When libcurl discards a connection there are two phases this may go through: "shutdown" and "closing". If a connection is aborted, the shutdown phase is skipped and it is closed right away. The connection filters attached to the connection implement the phases in their `do_shutdown()` and `do_close()` callbacks. Filters carry now a `shutdown` flags next to `connected` to keep track of the shutdown operation. Filters are shut down from top to bottom. If a filter is not connected, its shutdown is skipped. Notable filters that *do* something during shutdown are HTTP/2 and TLS. HTTP/2 sends the GOAWAY frame. TLS sends its close notify and expects to receive a close notify from the server. As sends and receives may EAGAIN on the network, a shutdown is often not successful right away and needs to poll the connection's socket(s). To facilitate this, such connections are placed on a new shutdown list inside the connection cache. Since managing this list requires the cooperation of a multi handle, only the connection cache belonging to a multi handle is used. If a connection was in another cache when being discarded, it is removed there and added to the multi's cache. If no multi handle is available at that time, the connection is shutdown and closed in a one-time, best-effort attempt. When a multi handle is destroyed, all connection still on the shutdown list are discarded with a final shutdown attempt and close. In curl debug builds, the environment variable `CURL_GRACEFUL_SHUTDOWN` can be set to make this graceful with a timeout in milliseconds given by the variable. The shutdown list is limited to the max number of connections configured for a multi cache. Set via CURLMOPT_MAX_TOTAL_CONNECTIONS. When the limit is reached, the oldest connection on the shutdown list is discarded. - In multi_wait() and multi_waitfds(), collect all connection caches involved (each transfer might carry its own) into a temporary list. Let each connection cache on the list contribute sockets and POLLIN/OUT events it's connections are waiting for. - in multi_perform() collect the connection caches the same way and let them peform their maintenance. This will make another non-blocking attempt to shutdown all connections on its shutdown list. - for event based multis (multi->socket_cb set), add the sockets and their poll events via the callback. When `multi_socket()` is invoked for a socket not known by an active transfer, forward this to the multi's cache for processing. On closing a connection, remove its socket(s) via the callback. TLS connection filters MUST NOT send close nofity messages in their `do_close()` implementation. The reason is that a TLS close notify signals a success. When a connection is aborted and skips its shutdown phase, the server needs to see a missing close notify to detect something has gone wrong. A graceful shutdown of FTP's data connection is performed implicitly before regarding the upload/download as complete and continuing on the control connection. For FTP without TLS, there is just the socket close happening. But with TLS, the sent/received close notify signals that the transfer is complete and healthy. Servers like `vsftpd` verify that and reject uploads without a TLS close notify. - added test_19_* for shutdown related tests - test_19_01 and test_19_02 test for TCP RST packets which happen without a graceful shutdown and should no longer appear otherwise. - add test_19_03 for handling shutdowns by the server - add test_19_04 for handling shutdowns by curl - add test_19_05 for event based shutdowny by server - add test_30_06/07 and test_31_06/07 for shutdown checks on FTP up- and downloads. Closes #13976
2024-06-19 18:40:06 +08:00
* progress the shutdown of connections in the queue.
*/
cpool: rename "connection cache/conncache" to "Connection Pools/cpool" This is a better match for what they do and the general "cpool" var/function prefix works well. The pool now handles very long hostnames correctly. The following changes have been made: * 'struct connectdata', e.g. connections, keep new members named `destination` and ' destination_len' that fully specifies interface+port+hostname of where the connection is going to. This is used in the pool for "bundling" of connections with the same destination. There is no limit on the length any more. * Locking: all locks are done inside conncache.c when calling into the pool and released on return. This eliminates hazards of the callers keeping track. * 'struct connectbundle' is now internal to the pool. It is no longer referenced by a connection. * 'bundle->multiuse' no longer exists. HTTP/2 and 3 and TLS filters no longer need to set it. Instead, the multi checks on leaving MSTATE_CONNECT or MSTATE_CONNECTING if the connection is now multiplexed and new, e.g. not conn->bits.reuse. In that case the processing of pending handles is triggered. * The pool's init is provided with a callback to invoke on all connections being discarded. This allows the cleanups in `Curl_disconnect` to run, wherever it is decided to retire a connection. * Several pool operations can now be fully done with one call. Pruning dead connections, upkeep and checks on pool limits can now directly discard connections and need no longer return those to the caller for doing that (as we have now the callback described above). * Finding a connection for reuse is now done via `Curl_cpool_find()` and the caller provides callbacks to evaluate the connection candidates. * The 'Curl_cpool_check_limits()' now directly uses the max values that may be set in the transfer's multi. No need to pass them around. Curl_multi_max_host_connections() and Curl_multi_max_total_connections() are gone. * Add method 'Curl_node_llist()' to get the llist a node is in. Used in cpool to verify connection are indeed in the list (or not in any list) as they need to. I left the conncache.[ch] as is for now and also did not touch the documentation. If we update that outside the feature window, we can do this in a separate PR. Multi-thread safety is not achieved by this PR, but since more details on how pools operate are now "internal" it is a better starting point to go for this in the future. Closes #14662
2024-08-23 19:58:41 +08:00
void Curl_cpool_multi_perform(struct Curl_multi *multi);
void Curl_cpool_multi_socket(struct Curl_multi *multi,
curl_socket_t s, int ev_bitmask);
lib: graceful connection shutdown When libcurl discards a connection there are two phases this may go through: "shutdown" and "closing". If a connection is aborted, the shutdown phase is skipped and it is closed right away. The connection filters attached to the connection implement the phases in their `do_shutdown()` and `do_close()` callbacks. Filters carry now a `shutdown` flags next to `connected` to keep track of the shutdown operation. Filters are shut down from top to bottom. If a filter is not connected, its shutdown is skipped. Notable filters that *do* something during shutdown are HTTP/2 and TLS. HTTP/2 sends the GOAWAY frame. TLS sends its close notify and expects to receive a close notify from the server. As sends and receives may EAGAIN on the network, a shutdown is often not successful right away and needs to poll the connection's socket(s). To facilitate this, such connections are placed on a new shutdown list inside the connection cache. Since managing this list requires the cooperation of a multi handle, only the connection cache belonging to a multi handle is used. If a connection was in another cache when being discarded, it is removed there and added to the multi's cache. If no multi handle is available at that time, the connection is shutdown and closed in a one-time, best-effort attempt. When a multi handle is destroyed, all connection still on the shutdown list are discarded with a final shutdown attempt and close. In curl debug builds, the environment variable `CURL_GRACEFUL_SHUTDOWN` can be set to make this graceful with a timeout in milliseconds given by the variable. The shutdown list is limited to the max number of connections configured for a multi cache. Set via CURLMOPT_MAX_TOTAL_CONNECTIONS. When the limit is reached, the oldest connection on the shutdown list is discarded. - In multi_wait() and multi_waitfds(), collect all connection caches involved (each transfer might carry its own) into a temporary list. Let each connection cache on the list contribute sockets and POLLIN/OUT events it's connections are waiting for. - in multi_perform() collect the connection caches the same way and let them peform their maintenance. This will make another non-blocking attempt to shutdown all connections on its shutdown list. - for event based multis (multi->socket_cb set), add the sockets and their poll events via the callback. When `multi_socket()` is invoked for a socket not known by an active transfer, forward this to the multi's cache for processing. On closing a connection, remove its socket(s) via the callback. TLS connection filters MUST NOT send close nofity messages in their `do_close()` implementation. The reason is that a TLS close notify signals a success. When a connection is aborted and skips its shutdown phase, the server needs to see a missing close notify to detect something has gone wrong. A graceful shutdown of FTP's data connection is performed implicitly before regarding the upload/download as complete and continuing on the control connection. For FTP without TLS, there is just the socket close happening. But with TLS, the sent/received close notify signals that the transfer is complete and healthy. Servers like `vsftpd` verify that and reject uploads without a TLS close notify. - added test_19_* for shutdown related tests - test_19_01 and test_19_02 test for TCP RST packets which happen without a graceful shutdown and should no longer appear otherwise. - add test_19_03 for handling shutdowns by the server - add test_19_04 for handling shutdowns by curl - add test_19_05 for event based shutdowny by server - add test_30_06/07 and test_31_06/07 for shutdown checks on FTP up- and downloads. Closes #13976
2024-06-19 18:40:06 +08:00
#endif /* HEADER_CURL_CONNCACHE_H */